Table 3 The structure and the performance statistics prediction.

From: A comparative analysis of artificial neural networks and wavelet hybrid approaches to long-term toxic heavy metal prediction

Metal

Model

Scenario

Input

R2

RMSE

As

BPNN

BPNN1

Fe, Flow, pH, WT, DO

0.550

0.383

BPNN2

Fe, pH, DO

0.415

0.376

BPNN3

Fe

0.442

0.163

NARX

NARX1

Fe, Flow, pH, WT, DO

0.537

0.512

NARX2

Fe, pH, DO

0.468

0.499

NARX3

Fe

0.279

0.255

WNN

WNN1

Fe, Flow, pH, WT, DO

0.122

0.279

WNN2

Fe, pH, DO

0.101

0.026

WNN3

Fe

0.439

0.178

WNARX

WNARX1

Fe, Flow, pH, WT, DO

0.321

0.475

WNARX2

Fe, pH, DO

0.631

0.300

WNARX3

Fe

0.335

0.278

Pb

BPNN

BPNN4

Fe, Flow, pH, NO3–N, EC

0.703

1.290

BPNN5

Fe, Flow

0.666

0.807

BPNN6

Fe

0.632

0.794

NARX

NARX4

Fe, Flow, pH, NO3–N, EC

0.621

1.006

NARX5

Fe, Flow

0.622

0.919

NARX6

Fe

0.611

0.777

WNN

WNN4

Fe, Flow, pH, NO3–N, EC

0.648

0.764

WNN5

Fe, Flow

0.691

0.714

WNN6

Fe

0.614

0.816

WNARX

WNARX4

Fe, Flow, pH, NO3–N, EC

0.013

3.306

WNARX5

Fe, Flow

0.039

1.085

WNARX6

Fe

0.602

0.761

Zn

BPNN

BPNN7

Fe, Flow, pH, NO3–N, EC

0.780

6.702

BPNN8

Fe, NO3–N

0.714

5.033

BPNN9

Fe

0.632

3.499

NARX

NARX7

Fe, Flow, pH, NO3–N, EC

0.385

9.280

NARX8

Fe, NO3–N

0.345

5.538

NARX9

Fe

0.575

4.067

WNN

WNN7

Fe, Flow, pH, NO3–N, EC

0.768

3.428

WNN8

Fe, NO3–N

0.700

3.425

WNN9

Fe

0.613

2.884

WNARX

WNARX7

Fe, Flow, pH, NO3–N, EC

0.034

10.188

WNARX8

Fe, NO3–N

0.006

13.727

WNARX9

Fe

0.637

3.407